Following the Money Muslim versus Muslim in Bosnia's Civil War Fotini Christia A largely understudied aspect of the 1992-95 Bosnian war-the intra-Muslim war in northwestern Bosnia-illustrates the importance of local elites and micro-level economic incentives in civil wars.' Though short factual references to the conflict abound, there exists no systematic explanation as to why the intra-Muslim civil war-which resulted in close to twice the number of battleground deaths as the Serb-Muslim conflict in that area-occurred in the first place.2 Why was there violent conflict within the Bosnian Muslim community at a time when its ethnic group was at war, facing intense violence and forcible expulsion by Serb and Croat forces? Primary source materials, including wartime documents as well as personal interviews with some of the war's leading actors, shed light on the reasons behind this abstruse internecine conflict. In civil wars where the broader conflict is cast in ethnic terms and the opportunity cost of fighting is high, that is, where GDP per capita is high, economic incentives can still seriously affect group behavior. In order to have this effect, there have to be high micro-level economic payoffs as well as local elites who can guarantee the survival oftheir constituents while providing access to these payoffs. Previous hypotheses fail to explain the puzzle of Bosnia's intra-Muslim war. However, simple tools from consumer choice theory can be utilized to lay out a theoretical conceptualization of the trade-off between ethnicity and economic payoffs during civil wars in the presence as well as the absence of strong local elites. Bosnia's intra-Muslim conflict illustrates this theoretical mechanism. Relevant Literature and Hypotheses The recent war in Bosnia-Herzegovina is often used as a paradigmatic case of ethnic conflict. It is traditionally cast in terms of three warring ethnic groups, the Bosnian Muslims, the Bosnian Serbs, and the Bosnian Croats, with accounts differingon the nature and extent of involvement of the latter's respective kin states of Serbia and Croatia. The Comparative Politics July 2008 overwhelming set of these works has focused on the macro-level ethnic cleavage rather than the possibly pertinent local cleavages.' Though many recognize the underlying role of a dilapidating economy and the decrease in foreign aid, very few works stray from the grievance path and look at nonethnic aspects as explanatory variable^.^ The intra-Muslim war in Bosnia is interesting in this regard because cxisting works on civil war can not readily explain it. The ethnic grievance school of thought can not account for the event since the conflict was intraethnic rather than interethni~.~And though the explanation that will be put forth in this article highlights the crucial role of economic incentives, dominant works in the greed-political economy of civil wars- school of thought can not justify the occurrence of such an event either. There are many rationales as to why and how ethnicity facilitates collective action. All lines of argument point to a clear pattern of preference for intraethnic unity rather than interethnic c~operation.~On a basic and rather intuitive level, the primordialist school of thought suggests that ethnic groups are important to individuals because they satisfy their need to belong. This hypothesis of in-group preference is empirically reinforced by experimental research in the social sciences which has noted a strong trend for cooperation between individuals within the same group rather than across groups.' In more instrumental terms, ethnicity can provide several formal or informal social institutions that encourage collective action through the opportunity for repeated interaction and information sharing. Repeated interaction leads to an understanding of reciprocity as well as reputation costs that facilitate in-group policing and sanctioning, rendering opportunistic defection costly.' Moreover, ethnicity provides a readily available shared code for communication and social interaction-be it a common language, religion, or shared customs and norms-that allows for easy coordination. In turn, it can be used as a standard metric of group inclusion and exclusion, providing a low cost mechanism for controlling government allocation of resources and for capturing rents.9 It is therefore puzzling to see the type of in-group defection that led to the intra- Muslim war in Bosnia, particularly at a time of interethnic war, when the Muslims were being massively ethnically cleansed by the Serbs and Croats. The present consensus among civil war scholars suggests that ethnicity becomes particularly salient once state- level institutions break down, inducing individuals to seek security in group solidarity and hardening what could have previously been fluid ethnic identities.I0 Even instrumental and constructivist understandings of ethnicity recognize that identity change involves certain costs that are arguably higher at times of conflict, leading to the expectation that individual preferences for coordination along ethnic lines would be even stronger in a conflict context." Though economic incentives carry a great degree of explanatory power, none of the hypotheses derived by the dominant works in the political economy of civil wars readily explains the intra-Muslim conflict. While the grievance school suggests that conflict can be explained by preferences, that it is the result of injustices and inequalities suffered Fontini Christia by certain groups that choose to rise up against their exploiters, the political economy school is more focused on opportunity: rebellion can at times be utilized as a source of profit, leading to civil war.I2 Poor states, proxied by low GDP per capita, are more likely to witness conflict because they constitute rich breeding grounds for rebel recruitment since the opportunity cost of fighting is low. Moreover, areas that are rich in natural resources, such as oil, diamonds, and drugs, have a higher risk of conflict, be it because these resources can finance rebel movements or, based on the degree of state dependency, because they cause reduced state capacity." While the per capita income in the Cazinska area of Bosnia that witnessed the intra-Muslim civil war was in relative terms higher than in other areas of Bosnia-that is, the opportunity cost of fighting was high-, and even though the area was not rich in any natural resources, it still saw conflict. Many of these studies have been partly inspired by international relations theories on interstate warfare and have employed international security paradigms to explain intrastate conflict.14 In line with that approach, they consider the main factions in the war as largely homogeneous units with a set of common preferences. Much like states, they treat ethnic groups, as well as different insurgent groups in conflict, in similar light: as unitary actors in a largely anarchic system. The relevant unit of analysis, therefore, remains aggregated, whether at the macro-level ethnic group, government, or insurgent group, with the elites and leaders of the group serving as the main actors behind the crucial decision making. A recent systematic critique of this approach suggests that it largely ignores internal group dynamics as well as individual incentives and preferences.I5 Such critiques have long existed in the field of anthropology, economics, and sociology but until recently were only seriously raised by area specialists in political science and were thus largely ignored. This recent line of work has directed interest to the micro level, compellingly highlighting the need to consider individual level incentives as a way to get an accurate grasp of the conflict processes at play.I6 In the civil war context, political and military elites are identified as macro-level actors while individuals-combatant~or civilian-are identified as micro-level actors. Even in the most democratic of contexts, however, the two are not always directly linked, and local elites serve as meso-level actors, that is, as intermediaries between the macro and micro levels. While identifying the set of macro and micro level rationales of action is undoubtedly important, it is equally important to focus on the set of macro and micro level interactions-the meso level-as a way to identify the relevant connections. More concretely, even in cases where the broader conflict is cast in largely macroethnic terms, local economic incentives can still prove important in the presence of charismatic local leadership (meso-level). In this case, ethnicity can be relegated to a cleavage of second- level importance. Economic payoffs can push people over the threshold where ethnicity matters, so much so that people are willing to fight against members of their own ethnic group. This effect can be at work only in the presence of strong local elites that can Comparative Politics Julv 2008 guarantee the survival of their people while providing access to micro-level economic payoffs. Theoretical Conceptualization A graphic representation from consumer choice theory can lay out a rationalist approach that informs the narrative of Bosnia's intra-Muslim conflict." Theoretically, this conflict could be simplistically conceptualized as a trade-off between two goods: economic resources and ethnic identity. One can picture it on a two-dimensional plane with economic resources as one of the dimensions (x-axis) and ethnic identity as the other (y-axis) (see Figure 1). In peacetime an individual's survival, understood as standard of living, calls for reliance on economic resources. One should be primarily interested in securing the highest degree of economic resources possible to improve one's living conditions, and unless one lives in a patronage society, ethnicity is not important in guaranteeing such resources. In peacetime, therefore, ethnicity is costly, that is, one is better off devoting most of one's endowment to acquiring economic resources. In economic and graphic terms, the budget constraint-the slope of the line that captures the limit on the consumption bundles that an individual can afford for a given income-is rather flat.